A grounded discussion of AI, software abundance, and the case for modern on-prem infrastructure.
O11ycast frames AI less as a stand-alone product category than as a force that could make software cheaper, more plentiful, and more locally useful.

How engineers see what their production code is really doing.
O11ycast, short for Observability Cast, is a technical conversation about how software teams find out what their code is really doing in production. Recent guests include Bryan Cantrill, Amarilis Campos of Nubank and Dr. Cat Hicks, and the question of how AI changes that work keeps coming up.
PRESS PLAY
O11ycast, spoken on air as Observability Cast, is a technical conversation about how software teams learn what their code is doing once it reaches production. Its stated aim is production systems that are more observable, more reliable and easier to maintain. That could pass for a tooling brief. The recent episodes keep pushing back on that reading. One puts the point flatly: 'This is not a tools problem. This is a systems problem.' The same conversation says the DevOps movement has been trying to build one feedback loop spanning developers and production, and that it has mostly failed. It adds that AI has shortened what counts as a fast feedback loop. The Nubank episode, with Amarilis Campos, calls observability a culture that is easy to undervalue when engineers are busy shipping new services. It also ties reliability to the whole customer journey, down to someone opening an app to send money to a parent or sister. AI comes up in most recent episodes, and the mood about it swings. In 'Software Is the Killer App for AI', with Bryan Cantrill, the falling cost of writing software is compared to what open source did, and a thousand kitchens in every home is described as good news for the fire department. The episode with Dr. Cat Hicks, a self-described psychologist for software teams, takes the other side: developers under heavy load while they adapt to a new technology, with little care for their psychological needs. It also argues that developer experience itself is poorly observed, a fitting complaint on a show named for observation. Then there is the hands-on material. Lada Kesseler walks through augmented coding patterns, including a 'context marker' that shows what an AI agent is holding in context. The same episode pictures an AI turned loose in Zork, a game from the 1980s, and asks what high-level views would reveal what it is up to. The conversation with Maddy Montaquila on .NET Aspire argues for telemetry by default, since stepping through a debugger stops working across fifty interacting microservices, and describes asking Honeycomb's MCP to read the telemetry and name the bug. There is some humor too, from a 'SpaghettiO' moment of panic to a detour into parenthood. Ken Rimple is heard in every recent introduction, Jessica Kerr in most, and Charity Majors, Austin Parker and Martin Thwaites in others. Every recent episode closes by inviting guest and topic suggestions, and Heavybit, an investor in developer-first startups, sponsors the series. The best place to start depends on the question a listener brings: the Cat Hicks episode for how AI tools affect teams, the Kesseler episode for practical habits with coding agents, and the Nubank conversation for reliability measured against credit card and lending transactions. Occasional mild profanity comes up in conversation.
Made for: Suits a listener who builds or runs production software and wants engineers talking plainly about telemetry, feedback loops and what AI-written code means for keeping systems understandable.
What sets it apart: It frames observability as a systems and culture problem rather than a tools problem. It also ranges widely in its guests, from a self-described psychologist for software teams to someone sharing coding patterns for AI agents.
Exploring the observability side of software development.
A perfect SLO, frustrated users: this is the argument your dashboard cannot settle.
Press play if you want
Short reviews from the PlayNext desk, based on the episodes we processed.
A grounded discussion of AI, software abundance, and the case for modern on-prem infrastructure.
O11ycast frames AI less as a stand-alone product category than as a force that could make software cheaper, more plentiful, and more locally useful.
A practical case for moving observability into the developer workflow as AI speeds up software work.
O11ycast brings Charity Majors, Ken Rimple, and Jessica Kerr into a focused discussion of why DevOps still often fails to connect development and production into a useful feedback loop.
A practical episode about making coding assistants' goals, context, and next steps easier to inspect.
Augmented coding patterns treats a coding assistant less like a magic autocomplete box and more like a working system that needs visibility. It opens with AI playing Zork, then turns that into a useful way to think about agents moving through a codebase: if you cannot see the…
A clear observability discussion that treats the customer journey as the real test of reliability.
Reliability is given a practical, customer-facing shape here: service health matters because someone is trying to complete a credit card transaction, make a lending transaction, or send money to family.
A careful discussion of why teams should listen to developers before AI expectations reshape software work.
Ollicast frames the AI-in-software-teams debate around the people being asked to absorb the change. With Kat Hicks, a psychologist for software teams and research architect, the episode looks at why developers may not always see the promised benefits of tools meant to reshape…
A grounded look at agents, learning, and the habits augmented coding may require.
O11ycast frames augmented coding less as a novelty than as a shift in the way software work gets shaped. Its strongest moments stay close to practice: an iOS app the guest does not yet know how to build, an agent treated as a power tool, and a useful reluctance to let "vibe…






Power Platform news with practical static.
["Brian Cantrell says software is becoming the killer app for AI, comparing the shift to open source and to his kitchen-fires metaphor about 'a thousand kitchens in every home.'", "He explains why Oxide Computer launched in 2019 around rack-mounted systems, API-driven infrastructure, and ownership instead of a cloud rental model.", "The episode also revisits how 'observability' emerged from `dtrace` work at Sun and spread through Twitter and related system teams.", "Cantrell argues that generative AI's real workload is CPU-bound software work, while the bigger macro story is still uncertain and often mislabeled as AI-driven layoffs."]
Charity Majors tells Ken Rimple and Jessica Kerr that metrics still fit infrastructure, but user-facing software needs structured logs and trace spans.
O11ycast episode #87, "Augmented Coding Patterns with Lada Kesseler," opens on the question of how you know what your coding assistant is doing.
Amarilis Campos of Nubank joins Observability Cast to discuss tracing, reliability, and observability for a bank that says it has more than 100 million customers across Brazil, Mexico, and Colombia.
ObservabilityCast #85 with Kat Hicks asks whether AI in software teams should be measured as one developer's speed or as a change to the code base and company.
["O11ycast Ep.
Dan Ravenstone says an observability engineer is not just doing DevOps or SRE work, because the job reaches product owners, customer success, and user experience.
O11ycast episode #82 follows Morgante Pell explaining how Grit started in 2022, before ChatGPT launched, to automate developer toil and technical debt.
ObservabilityCast's 'Observability 3.0-vNext-final-DRAFT' opens with the question, 'how much data that you emit do you ever look at?' Martin Thwaites and Charity Majors host Hazel Weakley and Matt Klein, who bring systems thinking, Envoy, AWS, Twitter, Lyft, and BitDrift into the same ROI debate.
O11ycast #80 centers on Kent Beck's argument for "augmented coding" over "vibe coding," and he says the world was ready for the concept.
Toli Apostolidis discusses Cinch's transition from container-based Kubernetes to serverless architecture and the operational challenges they faced.
Dmitry Filimonov discusses continuous profiling in production, addressing performance and storage concerns.
Aaron Aldrich discusses his role bridging skill gaps for Kubernetes adoption and shares insights from organizing DevOps Days Boston, comparing conference production monitoring to observability.
Nick Herring discusses his journey with technology decisions and cardinality problems in observability.
Jordan Simonovski discusses how Atlassian improves developer experience and observability practices at scale.